NEF

Compute Negative Eigen Fraction from PE dataset or mapping

    [F,R] = NEF(W)
    [F,R] = NEF(A)
    [F,R] = W*NEF
    [F,R] = A*NEF

Input
A	Pseudo-Euclidean dataset
W	Pseudo-Euclidean mapping

Output
F	Negative eigen-fraction
R	Negative eigen-ratio

Description

The negative eigen-fraction is a ratio expressing the sum of  magnitudes of all negative eigenvalues divided by the sum of  magnitudes of all eigenvalues. The necessary numbers are stored  in W = PE_EM(D) or in A = D*W, in which D is a symmetric dissimilarity  matrix.

The negative eigen-ratio is the ration of the magnitude of the lowest  negative eigenvalue to the largest positive eigenvalue.

See also

datasets, mappings, pe_em,